On the bias: Self-esteem biases across communication
channels during romantic couple conflict
Lauren Scissors, Darren Gergle
Northwestern University
2240 Campus Drive, Evanston, IL 60208 USA
{l-scissors, dgergle}@northwestern.edu
ABSTRACT
Are one’s individual biases stronger when mediated
communication is used? This paper examines the role of
self-esteem-related biases and communication channel
during romantic couple conflict. Romantic couples
communicated about a conflict either face-to-face (FtF) or
via instant messenger (IM). Results revealed that for people
with lower levels of self-esteem, their negative biases were
triggered when they communicated with their partners via
IM; people with lower levels of self-esteem had more
negative assessments of the conflict discussion and of the
impact of the discussion on the relationship when
communicating via IM than when communicating FtF. At a
theoretical level, this work deepens our understanding of
how individual difference variables like self-esteem impact
how individuals process information and communicate via
technology. At a practical level, findings suggest that the
use of mediated communication during conflict is more
harmful to certain individuals than to others.
Author Keywords
Conflict; romantic relationships; computer-mediated
communication; self-esteem
ACM Classification Keywords
H.5.m. Information interfaces and presentation (e.g., HCI):
Miscellaneous.
INTRODUCTION
Relational communication that was once reserved for faceto-face (FtF) encounters has spilled into the realm of
technology-mediated communication [47]. Individuals now
use technology to initiate, maintain, and even dissolve
relationships, and close relational ties are more likely to use
multiple communication channels to communicate than are
weak ties [16]. Yet, scholars are just beginning to uncover
how technology-mediated communication differs from FtF
communication in many relational contexts.
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that copies
bear this notice and the full citation on the first page. Copyrights for
components of this work owned by others than ACM must be honored.
Abstracting with credit is permitted. To copy otherwise, or republish, to
post on servers or to redistribute to lists, requires prior specific permission
and/or a fee. Request permissions from [email protected].
CSCW '16, February 27-March 02, 2016, San Francisco, CA, USA
© 2016 ACM. ISBN 978-1-4503-3592-8/16/02...$15.00
DOI: http://dx.doi.org/xx.xxxx/xxxxxxx.xxxxxxx
One context in which the use of communication
technologies is becoming more prevalent is interpersonal
conflict [48, 13, 37, 5]. Understanding how individuals
communicate during conflict is important since conflict is
inevitable in interpersonal relationships, and how partners
handle conflict can greatly impact relational outcomes (see
[39] for a review), as well as mental and physical wellbeing (see [24] for a review).
One group of people who are particularly drawn to using
technology to communicate with relational partners are
people with lower levels of self-esteem. Studies have
demonstrated that people with lower levels of self-esteem
(LSEs) prefer to use mediated rather than FtF
communication in a variety of contexts, likely because the
ambiguity of mediated communication can reduce facethreat (e.g, [21, 43]). Yet, it may be that the communication
channels LSEs prefer are actually ill-suited to their
communication and relational needs. For instance, one
study found that while LSEs were drawn to Facebook as a
way to connect with others through self-disclosure, their
disclosures were low in positivity and high in negativity,
which resulted in undesirable feedback from others [11]. In
the realm of close relationships, however, an open question
is: Do people with lower levels of self-esteem have more
negative relational experiences when using the channels of
communication that they prefer?
In romantic relationships, LSEs are biased toward negative
interpretations of their partners’ behavior and HSEs are
biased toward positive interpretations of their partners’
behavior [33, 34]. The goal of this work is to understand the
relationship between self-esteem and communication
channel in romantic couple conflict, a context in which selfesteem-related biases are particularly salient. The majority
of previous research in this area has involved hypothetical
situations, self-reports of past interactions, or controlled
experiments with strangers. The current work, however,
examines the roles of self-esteem and communication
channel in real-time communication among romantic
couples. LSEs are biased toward negative interpretations,
especially in ambiguous contexts, and how individuals
interpret their partners’ communication and behavior is
crucial to relational well-being [33]. Yet, we know little
about how communication channel can influence this
process for romantic couples. The current work aims to
contribute to our knowledge of how self-esteem biases are
enacted by examining those biases in a mediated
communication setting. By understanding the extent to
which self-esteem biases differ across communication
channels, we can see whether the communication channels
LSEs prefer are the ones that are actually harmful to their
relational well-being.
LITERATURE REVIEW
Self-Esteem in Romantic Relationships
Self-esteem represents how one feels about one’s self, or
one’s sense of worth, and is drawn in large part from our
perceptions of what others think of us. Individuals with low
self-esteem (LSEs) tend to have trouble in their romantic
relationships [33]. LSEs are biased toward negative
interpretations of their partners’ behavior and are sensitive
to feelings of threat and rejection. In addition, LSEs
dramatically underestimate how positively they are viewed
by their partners and have lower relational well-being [33]
and lower relational satisfaction [10, 33]. When faced with
relationship difficulties, LSEs search for information to
validate their thoughts that their partner does not care for
them and distance themselves from their partners as a way
to protect themselves from getting hurt [34].
In contrast to LSEs, individuals with high self-esteem
(HSEs) have positive biases when it comes to their partners,
and tend to interpret their partners’ communication and
behavior favorably. HSEs are also less sensitive to
relational threats, and respond to threats by emphasizing the
value of their relationship and drawing closer to their
partners [34]. The fact that they feel secure about
themselves likely helps HSEs to derogate relationship
threats or signs of rejection in order to focus on the positive
aspects of the relationship.
Yet, while LSEs pursue the goal of avoiding rejection, they
also pursue the competing goal of attaining closeness with a
relational partner [25]. The pursuit of these competing goals
is explained by the risk regulation system [32]. Murray and
colleagues explain that “the thoughts and behaviors that are
critical for establishing satisfying close connections with
others necessarily increase both the short-term risk of
rejection and long-term pain of rejection” ([32], p.642).
Since LSEs are so sensitive to rejection, their selfprotection goals often take precedence over their
relationship-enhancement goals.
Murray and colleagues have found that LSEs “read too
much into problems” with their partners, interpreting the
problems as signs of fading feelings or commitment [34].
Furthermore, Bellavia and Murray [31] found that LSEs
over interpret their dating partners’ negative moods,
internalizing responsibility and hurt feelings when the cause
of the mood was actually ambiguous. While LSEs may
exhibit negative interpretations because they are searching
for possible signs of rejection, LSEs may also be unwilling
to accept positive feedback from their partners since doing
so would not be consistent with their self-concepts [27].
This line of work suggests that LSEs are biased toward
negative interpretations, tend to overanalyze situations with
their partners, and tend to believe the worst in ambiguous
situations.
Overall, we know that LSEs are negatively biased toward
their partners’ communication and behavior, and that these
negative biases are particularly salient in threatening
contexts such as interpersonal conflict. Yet, it is possible
that the use of certain communication channels may
actually heighten these biases, since the amount and type of
social cues that help individuals interpret information differ
across channels.
Self-Esteem, Relationships, and Technology-mediated
Communication
When it comes to the domain of romantic couple conflict,
previous research has suggested that there are drawbacks of
using technology-mediated communication. For instance,
Burge and Tatar found that couples discussing a conflict via
mediated communication (phone and IM) may experience
significantly lower mood states than couples who
communicated FtF [5]. Another study found that when
individuals choose email over FtF to discuss a conflict, they
perceive their partners and themselves as more avoidant
[29]. In addition, Coyne and colleagues found that using
text messaging to discuss serious issues or broach a
potentially confrontational topic with one’s partner was
associated with increased levels of negative communication
in the relationship overall [8].
Using computer-mediated communication (CMC) is
typically thought of as a convenient way to avoid or ignore
one’s communication partner in face-threatening situations
like conflict (e.g., [28, 36]). This is likely because
affordances typical of channels like text messaging and IM
(e.g., reduced visibility, increased editability, reduced
amounts of nonverbal cues) create an environment that can
be more ambiguous than FtF communication. For instance,
if while texting one’s partner becomes unresponsive for
several minutes, it would be hard to know whether one is
being ignored or whether the partner is attending to
something important that suddenly came up. Or, if one sees
their partner is taking a long time to compose his/her
message, it would be hard to know whether the partner is
hesitant about responding or just taking great care to craft a
thoughtful response. This ambiguity can lead people to
draw on the available cues to form impressions [49], which
can result in exaggerated assessments of one’s
communication partner [38]. Walther describes this as the
“hyperpersonal model” of communication, which mainly
focuses on how individuals idealize or form positivelybiased impressions of their communication partners.
However, it is possible that in ambiguous contexts,
negatively biased impressions may form if people are
inclined to view their partners negatively.
Despite the previously described drawbacks of using CMC
during a conflict, there is also evidence that the use of
technology-mediated communication may actually be
helpful for individuals with insecurities. In a study of
individuals in dating relationships, increased frequency of
text messaging was associated with more positive
relationships for highly avoidant participants (those who
have a dismissive, independent attachment to relational
partners) but not for less avoidant participants [30]. This
finding suggests that text messaging affords avoidant
individuals a unique opportunity to connect with their
partners. In addition, greater use of social network sites was
associated with increased levels of intimacy and support in
relationships for anxiously attached participants (those who
have a worried, dependent attachment to relational partners)
but not for securely attached participants. Furthermore, [42]
found that individuals in romantic relationships seek out
mediated channels to better manage their emotions or to
find a less face-threatening way to offer a heartfelt apology
after a conflict. These findings suggest that the use of textbased communication may lead to positive relational
outcomes for people who feel insecure in their
relationships. This work, however, did not focus
specifically on the role self-esteem plays in mediated
communication.
In this study, we examine the extent to which self-esteem
biases manifest in mediated communication compared to
FtF communication. To do this, we brought romantic
couples to a lab, asked them to discuss a conflict either FtF
or via mediated communication, and then asked them to
provide assessments of the discussion.
HYPOTHESES
After couples discuss a conflict, how do self-esteem and
communication channel influence their assessments of that
discussion? We predict that self-esteem-related biases will
be stronger during mediated, text-based conversations than
during FtF conversations, since the increased ambiguity of
text-based communication may make it easier to perceive
what one is already inclined to perceive. More specifically,
we predict that during mediated conversations, LSEs’
negative biases will be heightened and, as a result, LSEs
will have more negative experiences with the interaction
than LSEs who have FtF conversations. In addition, we
predict that during mediated conversations, HSEs’ positive
biases will be heightened and, as a result, HSEs will have
more positive experiences with the interaction than HSEs
who have FtF conversations.
Overall Satisfaction with the Interaction
Based on prior literature about the role of self-esteem in
relationships, we predict:
H1a: LSEs will report lower ratings of satisfaction with the
interaction than will HSEs.
Based on prior literature, which suggests that the use of
mediated communication during conflict has negative
outcomes, we predict:
H1b:
Individuals
communicating
via
mediated
communication will report lower ratings of satisfaction
with the interaction than will individuals communicating
FtF.
And due to the biasing effects described earlier, we predict:
H1c: LSEs will report lower ratings of satisfaction with the
interaction when communicating with their partners via
mediated communication than when communicating FtF,
while HSEs will report higher ratings of satisfaction with
the interaction when communicating with their partners via
mediated communication than when communicating FtF.
Perceived Change in Relational Quality
In addition, individuals’ self-esteem-related biases may also
influence how they feel a particular conflict episode
impacts the state of their relationship, since LSEs often
devalue or seek to limit dependence on their relationships in
order to avoid further rejection (e.g., [32]). Previous
research has demonstrated that, in self-reports of past
relational conflicts, the use of mediated communication was
associated with reports of the conflict having more negative
impact on relational quality [43]. This same work did not
find a relationship between self-esteem and assessments of
a conflict’s impact on relational quality. It is possible,
however, that self-esteem biases are more easily detected in
the moment than via self-reports of a past event. Therefore,
we predict:
H2a: Lower levels of self-esteem will be associated with
more negative perceived change in relational quality.
H2b: Conflicts in mediated communication will be
associated with more negative perceived change in
relational quality than will FtF conflicts.
H2c: Lower levels of self-esteem will be associated with
more negative perceived change in relational quality for
conflicts in mediated communication than for FtF conflicts,
while higher levels of self-esteem will be associated with
more positive perceived change in relational quality for
conflicts in mediated vs. FtF communication.
METHOD
The study design draws on a number of instrumentations
from various studies of couple communication, including
studies of couple conflict (e.g., [1, 7, 14, 22, 44]). In these
studies, partners typically interact in some way (e.g.,
discuss a conflict topic) and provide feedback about the
experience. There is also a long history of studies that
examine aspects of interpersonal communication and
impression formation across different channels (e.g., [3, 4,
17]). In addition, Burge and Tatar [5] applied a version of
the couple conflict paradigm across three communication
channels (FtF, phone, and IM). The methodology employed
in this study draws on Burge and Tatar’s design (assigning
couples to discuss an argument in a particular
communication channel and then having couples report
individually on their experiences), on Afifi and colleagues’
[1] prompt for collecting potential conflict discussion
topics, and on various aspects of existing laboratory-based
work that examines interpersonal communication across
communication channels.
Participants
Participants (N = 176) were 88 romantic couples who were
either students or staff at a Midwestern U.S. university or
members of the surrounding community. Participants’ ages
ranged from 18 to 56 (M = 23.48, SD = 4.85). Each
participant received $25 for participating and the study
lasted about an hour.
Eighty-four couples were heterosexual and four couples
were same-sex couples (two were male-male couples and
two were female-female couples). Half of participants were
female and half were male. The majority of participants
were Caucasian (58%), with 18% Asian/Pacific Islander,
9% Hispanic/Latino, 8% African-American, and 7% Mixed
Race/Other (one participant did not report his/her race).
Relationship length ranged from 1 month to 12 years, with
an average of 23.7 months. The majority of couples
reported that they were “seriously dating” (72%), with 6%
“dating casually,” 7% “engaged,” 9% “married,” and 6%
“life partners.” Thirty-three percent of couples reported that
they lived together while 67% of couples reported that they
did not live together. Nine percent of couples described
themselves as being in a long-distance relationship, while
91% of couples described themselves as being in a nonlong-distance relationship.
Procedure
To investigate the extent to which self-esteem biases were
present across different communication channels, couples
were randomly assigned to one of two conditions: Instant
Messenger (IM) or Face-to-Face (FtF) communication.
Upon arriving at the lab, couples in both conditions were
consented together, and then they were put into two
separate rooms to fill out a series of questionnaire items.
First, participants filled out a series of questionnaire items
about themselves (including a measure of self-esteem) and
about their relationship (including their relationship status).
Participants were then provided with a short definition of
‘conflict’: “A conflict is an instance where you and your
partner are at odds with each other,” and each partner was
prompted to provide a list of four topics that tend to be
conflict inducing in their relationship (see [1] p.110 for
procedure details). Participants were told not to list any
topics they were uncomfortable discussing, that their
partner was also being asked to make a list, that their
partner would never see the participant’s list, and that the
researcher would compare their list with their partner’s in
order to pick a topic to discuss.
The researcher then compared the lists and searched for a
mutually listed topic. If multiple topics listed were mutual,
the researcher chose the first topic that was suggested in
both lists. If partners listed more than one mutual topic but
in different orders, one partner was randomly selected and
the topic that was higher on that partner’s list was selected.
If participants submitted lists that did not have any
overlapping issues, one partner was randomly selected and
the first issue on that partner’s list was discussed. All
participants wrote at least one topic.
After the topic was chosen, the researcher separately
informed each participant that they would be discussing this
topic with their partner in the next stage of the study. If the
partners mutually listed the topic, both wordings were
presented to the partners verbatim. Partners were told the
topic in their words first and in their partner’s words
second. If the partners did not list a mutual topic, the
researcher explained that this was the case and the
researcher randomly selected one of them and picked the
first item from that list.
At this point, the study diverged depending on the condition
to which couples were assigned. Accordingly, the FtF and
IM conditions are discussed separately below.
Face-to-Face Condition
For participants in the FtF condition, partners were brought
together in one room to discuss their topic. The researcher
then told participants that they would have eight minutes1 to
discuss the topic and that they would receive a one-minute
warning before time was up. They were told that they did
not have to use all of the allotted time and that either
partner could alert the researcher when they were done, if
they finished early. Then, the researcher started a video
recording and explained that once she left the room, the
participants could start talking. When the participants were
done talking, the researcher came back into the room,
stopped the recording, and escorted the partners to their
separate rooms. Then the researcher copied the video file to
each of the partner’s local computers and returned to the
participants’ rooms and told them that, for the next step,
they were going to review a video of the interaction they
just had with their partners. Participants then reviewed the
interaction and answered some questions about specific
things their partners said during the interaction. Lastly,
participants answered questions about their overall
satisfaction with the interaction and their perceived change
in relational quality resulting from the conflict discussion.
Finally, to restore their relationship to a more positive state,
participants were instructed to list three things that they like
about their relationship. Then they were debriefed.
1
This length of time is similar to other studies where couples
discuss a conflict (e.g., [2]).
2
Participants in the FtF condition had 8 minutes to talk while
participants in the mediated condition had 16 minutes to talk since
individuals tend to produce words at a lower rate when typing than
IM Condition
Participants in the IM condition remained in their separate
rooms after they were told which topics they would be
discussing. The researcher then opened up an Instant
Messaging client on each of the participants’ computers.
Partner A’s screen name was “Participant A” and Partner
B’s screen name was “Participant B.” (Participants also had
a sticky note on the bottom of their monitors to remind
them of which letter they were assigned to). The researcher
separately explained to participants that they would have 16
minutes2 to discuss the topic with their partners, and that
they would receive a two-minute warning when time was
almost finished. As in the FtF condition, participants were
told that they did not have to use all of the allotted time and
that either partner could alert the researcher when they were
done, if they finished early.
Once the time was up or once participants were finished,
participants reviewed a PDF transcript of the conversation
they just had with their partners and answered some
questions about specific things their partners said during the
interaction. As in the FtF condition, participants in the IM
condition also answered questions about their overall
satisfaction with the interaction and their perceived change
in relational quality resulting from the conflict discussion.
They also completed the relationship-restoring task and
were debriefed.
Full IRB approval was obtained for this study from the
authors’ institution and all data was anonymized.
MEASURES
Items were measured on a 7-point Likert scale (1 = strongly
disagree, 7 = strongly agree) unless otherwise indicated.
[12]. The trait of self-esteem has been found to have similar
effects in romantic relationships as the trait of attachment
anxiety [6]. In other words, individuals with low selfesteem (LSEs) act in similar ways to anxiously attached
individuals, and individuals with high self-esteem (HSEs)
act in similar ways to securely attached individuals. In
addition, securely attached individuals tend to have higher
levels of self-esteem than do individuals with less secure
attachment styles (see [26] for a brief review). Yet,
attachment anxiety and self-esteem, while related, are
typically treated as independent constructs (e.g., [15]).
Partner-specific attachment anxiety was measured using a
modified portion of the Experiences in Close Relationships
Scale (ECR)-Short Form [53] which contained a six-item
measure of attachment anxiety (M = 3.12, SD = .99, α =
.68). (Example item: “I need a lot of reassurance that I am
loved by my partner.”)
Relationship Satisfaction
Participants completed a 10-item relationship satisfaction
scale [41], which included items like “My partner fulfills
my needs for intimacy,” and “I feel satisfied with our
relationship” (M = 6.23, SD = .69, α = .90).
Dependent Variables
Overall Satisfaction with the Interaction
Participants filled out a one-item measure from the Iowa
Communication Record [9]: “Please indicate the extent to
which you came away satisfied with the interaction” (1 =
not at all satisfied, 7 = very satisfied), (M = 5.38, SD =
1.45).
Perceived Change in Relational Quality
Independent Variables
Communication Channel
Half of the couples in the study communicated via instant
messenger and half communicated face-to-face.
Self-Esteem
Participants completed the 10-item Rosenberg self-esteem
scale [40], which included items like “On the whole, I am
satisfied with myself,” and “I take a positive attitude toward
myself” (M = 5.47, SD = .95, α = .89).
Partner-Specific Attachment Anxiety
In addition to self-esteem, attachment anxiety is a construct
that reflects the feelings of insecurity in romantic
relationships. Anxiously attached (vs. securely attached)
individuals are insecure about their relationship partners,
and search for information to corroborate these insecurities
2
Participants in the FtF condition had 8 minutes to talk while
participants in the mediated condition had 16 minutes to talk since
individuals tend to produce words at a lower rate when typing than
when speaking (see [19], [18] for examples).
Participants completed a 3-item scale measuring the extent
to which the conflict discussion changed or impacted the
quality of the relationship [20]: “After your discussion
today, did your relationship become …?” “more distant (1)
or closer (7)”, “weaker (1) or stronger (7)”, “more sad (1)
or happier (7)” (M = 5.36, SD = 1.13, α = .95).
ANALYSIS
Given that participants are part of a couple, it is possible
that participants’ answers to questions about overall
satisfaction with the interaction and perceived change in
relationship may not be wholly independent of their
partners’ answers. To determine whether a significant
proportion of the variance in the dependent variables are
accounted for by the couple, which we refer to as the dyad,
a series of intraclass correlations were conducted using the
double entry method [23]. There was a positive, significant
intraclass correlation between one’s overall satisfaction
with the interaction and one’s partner’s overall satisfaction
(r = .33, p < .01). The intraclass correlation between one’s
ratings of the conflict’s impact on relational quality and
one’s partner’s ratings of the conflict’s impact on relational
quality was .23 (p < .01). The size of these correlations
suggests that, in the cases where correlations were
significant, there is a lack of independence between the
measures taken within a dyad.
To control for the effect of the dyad and for the role of
variables that are known to be correlated with self-esteem
(partner-specific
attachment
anxiety,
relationship
satisfaction), two mixed effects hierarchical regression
models were conducted to examine the role of self-esteem,
communication channel (condition), and the interaction
effect between self-esteem and channel (independent
variables) on overall satisfaction with the interaction, and
perceived change in relational quality. In these models, the
individual, nested in the pair, was included as a random
effect. Relationship satisfaction and partner-specific
attachment anxiety were included as control variables.
Variable
Intercept
Relationship Satisfaction
Partner-specific Attachment Anxiety
Self-Esteem
Channel [FtF]
Self-esteem × Channel [FtF]
B
.34
.57
.09
.22
−.04
−.21
SE
p
1.40 .81
.16 .00**
.12 .43
.12 .06^
.11 .69
.11 .06^
2
R = .12
** p <.01, * p < .05, ^ p < .10
Effect of Dyad: Variance = 1.94, SE = .21, Upper CI = 1.56,
Lower CI = 2.38
Table 1. Model predicting ratings of overall satisfaction with
the interaction.
RESULTS
Overall Satisfaction with the Interaction
Participants rated the extent to which they were satisfied
with the interaction (the conflict discussion they had with
their partners). Results (presented in Table 1) revealed a
main effect of self-esteem at the trend level (H1a
supported), such that higher levels of self-esteem were
associated with higher ratings of satisfaction with the
interaction. Results also revealed a significant interaction
effect (at the trend level) of self-esteem × channel such that
higher levels of self-esteem were associated with higher
ratings of overall satisfaction with the interaction (i.e., the
conversation) in IM than they were in FtF (H1c supported).
Or, put another way, lower levels of self-esteem were
associated with lower ratings of overall satisfaction with the
interaction in IM than they were in FtF. Examination of the
slope of the two lines (see Figure 1), reveals that while
ratings of overall satisfaction with the interaction vary by
levels of self-esteem for participants in the IM condition (B
= .43, t = 2.72, p < .01), self-esteem does not have a
significant effect on ratings of overall satisfaction with the
interaction for participants in the FtF condition (B = .01, t =
.07, p = .95). This suggests that self- esteem biases were not
present in the FtF condition, but were present in the IM
condition.
There was no main effect of channel (H1b not supported).
As for the control variables, higher levels of relationship
satisfaction were associated with higher ratings of overall
satisfaction with the interaction, but partner-specific
attachment anxiety was not a significant predictor of
satisfaction with the interaction.
Perceived Change in Relational Quality
In terms of participants’ reports of perceived change in
relational quality after the conflict discussion, there was no
main effect of self-esteem (H2a not supported) or channel
(H2b not supported) (See Table 2). However, there was an
interaction effect such that for LSEs, perceptions of change
in relational quality were more negative for those who
Figure 1. Self-esteem x channel (FtF vs. IM) predicting ratings
of overall satisfaction with the interaction.
communicated via IM than they were for those who
communicated FtF, and for HSEs, perceptions of change in
relational quality were more positive for those who
communicated FtF. The slope of the line in the IM
condition is statistically significant (B = .27, t = 2.24, p <
.05), while the slope of the line in the FtF condition is not
statistically significant (B = −.11, t = -.87, p = .38) (See
Figure 2). This suggests that self-esteem biases were not
present in the FtF condition, but were present in the IM
condition. Higher levels of relationship satisfaction were
also significantly associated with more positive perceptions
of change in relational quality. In other words, individuals
who were more satisfied in their relationships rated the
conflict’s impact on relational quality as more positive than
did individuals who were less satisfied in their
relationships. Partner-specific attachment anxiety was not a
significant predictor of perceived change in relational
quality.
Variable
Intercept
Relationship Satisfaction
Partner-specific Attachment Anxiety
Self-Esteem
Channel [FtF]
Self-esteem × Channel [FtF]
B
1.28
.53
.11
.08
−.05
−.19
SE
p
1.08 .24
.12 .00**
.09 .24
.09 .37
.08 .55
.09 .05*
2
R = .13
** p <.01, * p < .05, ^ p < .10
Effect of Dyad: Variance = 1.15, SE = .13, Upper CI = 1.44,
Lower CI = .94
Table 2. Model predicting ratings of perceived change in
relational quality.
Figure 2. Self-esteem x channel (FtF vs. IM) predicting ratings
of perceived change in relational quality.
DISCUSSION
Previous research has found that people with lower levels
of self-esteem prefer to use mediated communication across
a variety of relational contexts. The current research
suggests that the text-based mediated channels LSEs desire
may actually be detrimental to their relationships. Higher
levels of self-esteem were associated with higher ratings of
overall satisfaction with the interaction in IM than they
were in FtF, and lower levels of self-esteem were associated
with lower ratings of overall satisfaction with the
interaction in IM than they were in FtF. The same results
were found for ratings of perceived change in relational
quality. LSEs reported that the conflict had a more negative
impact on relational quality after communicating with their
partners via IM than they did after communicating with
their partners FtF. These findings provide evidence for the
fact that individuals’ self-esteem related biases were
exhibited when communicating via IM.
The fact that LSEs have different experiences with their
romantic partners based on which communication channel
they use is of crucial importance, given the popularity of
technology-mediated channels like text-messaging and
SNSs, and the fact that several studies have now shown that
individuals use technology-mediated communication during
a conflict with their romantic partners. If using a text-based
technology like IM during a conflict becomes the norm,
rather than the exception, LSEs may be at an even further
disadvantage than they already are given their biases. One
solution may be to suggest using multiple channels during a
conflict to reap the benefits that each channel uniquely
provides. For instance, [42] found that CMC can be helpful
during a conflict as a way to ease into a future FtF-based
conflict discussion and that different channels can
compliment each other in different ways. Individuals could
use a variety of communication channels across a conflict
discussion to best handle their insecurity-related biases and
their relational goals.
In addition, these findings may have implications for people
in long-distance relationships (LDRs) since they have less
FtF communication with their partners (e.g., [46]). Previous
research has demonstrated that LDRs are different from
non-LDRs in several ways. For instance, people in LDRs
are more likely to avoid conflict, engage in selective
positive self-presentation [45], and idealize their partners
relative to people in non-LDRs [46]. However, for people
in LDRs with low self-esteem, the reliance on technologymediated communication may actually be a detriment to
their relationships. Yet, people in LDRs may have more
channel choices than just FtF and IM; future research could
investigate the extent to which our findings hold for people
in LDRs and across different communication channels like
video chat, which has benefits and drawbacks for people in
LDRs that are unique from other channels [35].
Results were not entirely in line, however, with what we
hypothesized. Self-esteem-related biases were not exhibited
in the FtF condition. The lack of self-esteem-related biases
in the FtF condition may be due to sentiment override [54],
such that participants’ high levels of relationship
satisfaction were stronger predictors of ratings of overall
satisfaction with the interaction than were levels of selfesteem. However, it may be that the ambiguity of IM was
enough to trigger participants’ self-esteem biases such that
these biases influenced ratings of overall satisfaction with
the interaction above and beyond the influence of
relationship satisfaction. Given that more elements of the
interaction that typically indicate meaning (e.g., partner’s
facial expressions, tone of voice) were unknown in the IM
condition, it is possible that this ambiguity triggered LSEs’
instincts to question their partners’ acceptance and devalue
their relationships and HSEs’ instincts to draw closer to
their partners and enhance their relationships.
Another possible explanation is that LSEs and HSEs
actually experience similar levels of satisfaction with the
interaction in IM but that, because LSEs have a stronger
preference for, and potentially higher expectations of, using
CMC during a conflict with their partner (as found in [43]),
when their experiences do not live up to their expectations,
they experience cognitive dissonance and thus lower levels
of satisfaction with the interaction. Alternatively, HSEs
may actually reaffirm their relationship after an IM conflict
discussion (through higher ratings of satisfaction with the
interaction), precisely because the ambiguity of IM made
them have a more negative experience than they might have
expected. This supposition is somewhat complementary to
the hypothesized relationship between self-esteem and
channel (IM’s ambiguity allows HSEs more room to project
their positive biases), since in both cases, HSEs project
positive biases. But, in this alternative explanation, the
positive biases emerge to counteract any negative feelings
resulting from communicating via IM, as opposed to the
hypothesized relationship, which is that HSEs’ already
positive biases would be even more positive given more
room for interpretation.
Earlier, we discussed Walther’s hyperpersonal model of
communication, which posits that in text-based mediated
communication, people may form more exaggerated
impressions given that there are fewer relational cues to
draw on than there are in FtF communication. The current
findings suggest, however, that the hyperpersonal model of
communication does not fully account for the individualdifference level traits that influence the processing of social
information in mediated communication. Individuals make
use of the cues available to form impressions of their
communication partners, but this finding suggests that
individuals enter into conversations with pre-set biases that
also influence the ways in which they form impressions.
This finding then suggests that future work investigating
theories of mediated communication should not neglect the
biases that individuals bring to an interaction, whether they
be related to self-esteem or other traits. Furthermore, while
the hyperpersonal model allows for the possibility that
hyperpersonal communication can be negative, this part of
the model is not fully theorized. Future research should
further investigate how hyperpersonal communication can
lead to interactions that are intensified in a negative
manner, and how individual difference biases like selfesteem or personality influence this communication
process.
External Validity
Many of the conflict topics discussed by participants have
been identified as relational conflict issues in previous work
(see [37] for a review). Topics discussed included:
communication styles, time management, responsibilities,
one partner’s personal problem/shortcoming, past
relationships, jealousy, trust, lying, sex life, relationship
commitment level, living situation, and plans for the
future/life choices. Also, it is possible that participants held
back during their discussions due to the fact that they were
in a laboratory setting. However, we asked participants
“How realistic was your conversation with your partner” (1
= not realistic at all, 7 = very realistic). The average score
was 5.74, suggesting that there is some external validity to
the study.
Limitations and Future Work
An important caveat to this work is that, while ‘LSEs’ and
‘lower levels of self-esteem’ are used somewhat
interchangeably, self-esteem was measured as a continuous
variable and not as a categorical variable (e.g., low selfesteem vs. high self-esteem), suggesting that analyses did
not truly compare individuals with low vs. high self-esteem.
Future work should examine how people who are much
lower in self-esteem make assessments of conflict
discussions with their partners across communication
channels.
In addition, since only nine couples in the study were in
long-distance relationships, we do not have enough data to
analyze whether our findings may differ for people in longdistance vs. non-long-distance relationships. Since couples
in long-distance relationships may use technology to
communicate in unique ways, future work should examine
whether these findings hold for people in long-distance
relationships.
Lastly, this work focuses on one form of CMC – IM.
Results may differ across other forms of mediated
communication, such as text messaging, email, and video
chat, which are all distinct from IM in several ways. Also,
while this study examined a conflict discussion that
occurred in one communication channel, previous work
suggests that couples use multiple channels to communicate
during a conflict [42, 43]. Future work should focus on
examining the extent to which self-esteem biases are
enacted across other forms of CMC, as well as on how
experiences communicating through one channel might
influence perceptions of the interaction in another channel.
CONCLUSION
Overall, findings suggest that individuals’ self-esteem
biases are stronger in CMC than in FtF when it comes to
their satisfaction with the interaction and assessments of the
conflict’s impact on relational quality. This suggests that
while people with lower levels of self-esteem prefer
mediated communication, their preferred channels may not
be best suited for their biases and lead to worse
interpersonal outcomes.
ACKNOWLEDGEMENTS
We thank Danielle Pierre for her help with data collection,
members of the CollabLab for their support, and the
anonymous reviewers for their suggestions and insights,
which greatly improved this work.
REFERENCES
1. Tamara D. Afifi, Andrea Joseph, and Desiree Aldeis.
2012. The "standards for openness hypothesis": Why
women find (conflict) avoidance more dissatisfying than
men. J. of Social and Personal Relationships, 29(1),
102-125.
2. Christin E. Bates and Jennifer A. Samp. 2011.
Examining the effects of planning and empathic
accuracy on communication in relational and
nonrelational conflict interactions. Communication
Studies, 62(2), 207-223.
3. Nathan Bos. Judy Olson, Darren Gergle, Gary Olson,
and Zach Wright. 2002. Effects of four computermediated communications channels on trust
development. CHI 2002, 135-140.
4. Eliane M. Boucher, Jeffrey T. Hancock, and Philip J.
Dunham. 2008. Interpersonal Sensitivity in ComputerMediated and Face-to-Face Conversations. Media
Psychology, 11, 235-258.
5. Jamika D. Burge and Deborah Tatar. 2009. Affect and
dyads: Conflict across different technological media. In
S. Harrison (Ed.), Media Space 20+ Years of Mediated
Life (pp. 123-144): Springer.
6. Lorne Campbell, Jeffry A. Simpson, Jennifer Boldry,
and Deborah A. 2005. Perceptions of conflict and
support in romantic relationships: The role of
attachment anxiety. J. of Personality and Social
Psychology, 88(3), 510-513.
7. Nancy L. Collins and Brooke C. Feeney. 2004. Working
models of attachment shape perceptions of social
support: Evidence from experimental and observational
studies. J. of Personality and Social Psychology, 87(3),
363-383.
8. Sarah M. Coyne, Laura Stockdale, Dean Busby,
Bethany Iverson, and David M. Grant. 2011. "I luv u :)":
A descriptive study of the media use of individuals in
romantic relationships. Family Relations, 60, 150-162.
9. Steve Duck, Deborah Rutt, Margaret Hoy Hurst, and
Heather Strejc. 1991. Some evident truths about
conversations in everyday relationships: All
communications are not created equal. Human
Communication Research, 18(2), 228-267.
10. Frank D. Fincham and Thomas N. Bradbury. 1993.
Marital satisfaction, depression, and attributions: A
longitudinal analysis. Journal of Personality and Social
Psychology, 64, 442-452.
11. Amanda L. Forest and Joanne V. Wood. 2012. When
social networking is not working: Individuals with low
self-esteem recognize but do not reap the benefits of
self-disclosure on Facebook. Psychological Science,
23(3), 295-302.
12. Chris R. Fraley and Phillip R. Shaver. 2000. Adult
romantic attachment: Theoretical developments,
emerging controversies, and unanswered questions.
Review of General Psychology, 4, 132-154.
13. Brandi N. Frisby and David Westerman. 2010. Rational
actors: Channel selection and rational choices in
romantic conflict episodes. Journal of Social and
Personal Relationships, 27(7), 970-981.
14. John M. Gottman, James Coan, Sybil Carrere, and
Catherine Swanson. 1998. Predicting marital happiness
and stability from newlywed interactions. Journal of
Marriage and Family, 60(1).
15. Joshua Hart, Phillip R. Shaver, and Jamie L Goldenberg.
2005. Attachment, self-esteem, worldviews, and terror
management: Evidence for a Tripartite security system.
Journal of Personality and Social Psychology, 88(6),
999-1013.
16. Caroline Haythornthwaite. 2005. Social networks and
internet connectivity effects. Information,
Communication & Society, 8(2), 125-147.
17. Andrew C. High and Scott E. Caplan. 2009. Social
anxiety and computer-mediated communication during
initial interactions: Implications for the hyperpersonal
perspective. Computers in Human Behavior, 25, 475482.
18. Elizabeth V. Hobman, Prashant Bordia, Bernd Irmer,
and Artemis Chang. 2002. The expression of conflict in
computer-mediated and face-to-face groups. Small
Group Research, 33(4), 439-465.
19. L. Crystal Jiang, Natalie N. Bazarova, and Jeffrey T.
Hancock. 2013. From perception to behavior:
Disclosure reciprocity and the intensification of
intimacy in computer-mediated communication.
Communication Research, 40(1), 125-143.
20. Kristen L. Johnson and Michael E. Roloff. 2000.
Correlates of the perceived resolvability and relational
consequences of serial arguing in dating relationships:
Argumentative features and the use of coping strategies.
Journal of Social and Personal Relationships, 17(4-5),
676-686.
21. Adam N. Joinson. 2004. Self-esteem, interpersonal risk,
and preference for e-mail to face-to-face
communication. CyberPsychology & Behavior, 7(4),
479-485.
22. K. Laura Keck and Jennifer A. Samp. 2007. The
dynamic nature of goals and message production as
revealed in a sequential analysis of conflict interactions.
Human Communication Research, 33, 27-47.
23. David A. Kenny, Deborah A. Kashy, William L. Cook.
2006. Dyadic Data Analysis. New York, NY: The
Guilford Press.
24. Janice K. Kiecolt-Glaser and Tamara L. Newton. 2001.
Marriage and health: His and hers. Psychological
Bulletin, 127(4), 472-503.
25. Mark R. Leary and Roy F. Baumeister. 2000. The nature
and function of self-esteem: Sociometer theory In M. P.
Zanna (Ed.), Advances in experimental psychology (pp.
1-62). San Diego, CA: Academic Press.
26. Geoff MacDonald and Mark R. Leary. 2012. Individual
differences in self-esteem. In M. R. Leary & J. P.
Tangney (Eds.), Handbook of Self and Identity. New
York, NY: The Guilford Press.
27. Denise C. Marigold, John G. Holmes, and Michael
Ross. 2007. More than words: Reframing compliments
from romantic partners fosters security in low selfesteem individuals. Journal of Personality and Social
Psychology, 92(2), 232-248.
28. Erich R. Merkle and Rhonda A. Richardson. 2000.
Digital dating and virtual relating: Conceptualizing
computer mediated romantic relationships. Family
Relations, 49, 187-192.
29. Monique M. Mitchell, Caroline T. Rankin, Scott C.
D'Urso, and Patty Malone. 2002. The impact of intention
to handle conflict face-to-face or via e-mail on
perceived conflict resolution style. Paper presented at
ICA 2002.
30. Jennifer N. Morey, Amy L. Gentzler, Brian Creasy, Ann
M. Oberhauser, and David Westerman. 2013. Young
adults’ use of communication technology within their
romantic relationships and associations with attachment
style. Computers in Human Behavior, 29, 1771-1778.
31. Sandra L. Murray, Gina M. Bellavia, Paul Rose, and
Dale W. Griffin. 2003. One hurt, twice hurtful: How
perceived regard regulates daily martial interaction.
Journal of Personality and Social Psychology, 84(1),
126-147.
32. Sandra L. Murray, John G. Holmes, Nancy L. Collins, .
2006. Optimizing assurance: The risk regulation system
in relationships. Psychological Bulletin, 132(5), 641666.
33. Sandra L. Murray, John G. Holmes, and Dale W.
Griffin. 2000. Self-esteem and the quest for felt security:
How perceived regard regulates attachment processes.
Journal of Personality and Social Psychology, 78(3),
478-498.
34. Sandra L. Murray, Paul Rose, Gina M. Bellavia, John G.
Holmes, and Anna G. Kusche. 2002. When rejection
stings: How self-esteem constrains relationshipenhancement processes. Interpersonal Relations and
Group Processes, 83(3), 556-573.
35. Carman Neustaedter and Saul Greenberg. (2012).
Intimacy in long-distance relationships over video chat.
CHI 2012753-762.
36. Patrick B. O'Sullivan. 2000. What you don't know won't
hurt me: Impression management functions of
communication channels in relationships. Human
Communication Research, 26(3), 403-431.
37. Ofcom. (2013). Communications Market Report
http://stakeholders.ofcom.org.uk/binaries/research/cmr/c
mr13/2013_UK_CMR.pdf. .
38. Artemio Ramirez, Jr., Joseph B. Walther, Judee K.
Burgoon, and Michael Sunnafrank. 2002. Informationseeking strategies, uncertainty, and computer-mediated
communication: Toward a conceptual model. Human
Communication Research, 28(2), 213-228.
39. Michael Roloff and Kari P. Soule. 2002. Interpersonal
Conflict: A review. In M. L. Knapp & J. A. Daly (Eds.),
Handbook of Interpersonal Communication, 3rd edition.
Thousand Oaks, CA: Sage.
40. Morris Rosenberg. 1965. Society and the adolescent
self-image. Princeton, NJ: Princeton University Press.
41. Caryl E. Rusbult, John M. Martz, and Christopher R.
Agnew. 1998. The investment model scale: Measuring
commitment level, satisfaction level, quality of
alternatives, and investment size. Personal
Relationships, 5, 357-391.
42. Lauren E. Scissors and Darren Gergle. 2013. “Back and
forth, back and forth”: Channel switching in romantic
couple conflict. CSCW 2013, 237-248.
43. Lauren Scissors, Michael E. Roloff, and Darren Gergle.
2014. Room for interpretation: The role of self-esteem
and CMC in romantic couple conflict. CHI 2014, 39533962.
44. Alan Sillars, Traci Smith, and Ascan Koerner. 2010.
Misattributions contributing to empathic (in) accuracy
during parent-adolescent conflict discussions. Journal of
Social and Personal Relationships, 27(6), 727-747.
45. Laura Stafford. 2010. Geographic distance and
communication during courtship. Communication
Research, 37(2), 275-297.
46. Laura Stafford and Andy J. Merolla. 2007. Idealization,
reunions, and stability in long-distance dating
relationships. Journal of Social and Personal
Relationships, 24(1), 37-54.
47. Stephanie Tom Tong and Joseph B. Walther. 2011.
Relational maintenance and computer-mediated
communication. In K. B. Wright & L. M. Webb (Eds.),
Computer-mediated communication in personal
relationships (pp. 98-119). New York: Peter Lang.
48. Monique Mitchell Turner, Scott C. D'Urso, Caroline T.
Rankin, and Patty Callish Malone. 2003. Cognition
during conflict: The impact of communication medium
and perceived intentions on attributions. Paper
presented at ICA 2003.
49. Joseph B. Walther. 1992. Interpersonal effects in
computer-mediated interaction: A relational perspective.
Communication Research, 19(1), 52-90.
50. Joseph B. Walther. 1996. Computer-mediated
communication: Impersonal, interpersonal, and
hyperpersonal interaction. Communication Research,
23(1), 3-43.
51. Joseph B. Walther. 2011. Theories of computermediated communication and interpersonal relations. In
M. L. Knapp & J. A. Daly (Eds.), The handbook of
interpersonal communication (4th ed., pp. 443-479).
Thousand Oaks, CA: Sage.
52. Joseph B. Walther, Tracy Loh,and Laura Granka. 2005.
Let me count the ways: The interchange of verbal and
nonverbal cues in computer-mediated and face-to-face
affinity. J. of Language and Social Psychology, 24(1),
36-65.
53. Meifen Wei, Daniel W. Russell, Brent Mallinckrodt, and
David L. Vogel. 2007. The experiences in close
relationship scale (ECR)-short form: Reliability,
validity, and factor structure. J. of Personality
Assessment, 88, 187-204.
54. Robert L. Weiss. 1980. Strategic behavioral marital
therapy: Toward a model for assessment and
intervention. In J. P. Vincent (Ed.), Advances in family
intervention, assessment, and theory (Vol. 1, pp. 229271). Greenwich, CT: JAI Press.